your indexed records for @article{Douglass_2011, doi = {10.1007/s12237-010-9356-4}, url = {https://doi.org/10.1007%2Fs12237-010-9356-4}, year = 2011, month = {jan}, publisher = {Springer Nature}, volume = {34}, number = {4}, pages = {701--711}, author = {James G. Douglass and J. Emmett Duffy and Elizabeth A. Canuel}, title = {Food Web Structure in a Chesapeake Bay Eelgrass Bed as Determined through Gut Contents and 13C and 15N Isotope Analysis}, journal = {Estuaries and Coasts}} #1
Comments
|
hey @ddkapan - thanks for reporting this suspicious interaction. I believe I traced the suspicious interactions to network 113 in mangal.io - it appears that their herbivory and detrivore interactions were flipped. from https://mangal.io/api/v2/interaction?network_id=1113 - it appears in this interaction the plant was node_from and the plant eater was node_to , whereas other interactions of same type in different networks in mangal have the So, as far as I can tell, this is a likely transcription error in mangal. |
|
Perhaps you can contact the maintainers of mangal (@tpoisot or @SteveViss) about this suspicious interaction. |
|
We're aware of it and it'll be fixed soon - @ddkapan GLOBI is (in my understanding) not a data source, so don't hesitate to reach out if you find more suspicious interactions. Most of what is in mangal has been curated from the supp. mat. of papers, but we may have made a few mistakes. |
|
Thanks for responding so quickly. I am glad that GloBI was able to facilitate this exchange. Please holler as soon as you fix is so we can confirm and close this issue. |
|
Just for your information, it appears there's at least on existing Douglass 2011 interaction reported via mangal that has plants eating Gammarus sp. See attached screenshot related to https://www.globalbioticinteractions.org/?interactionType=eats&sourceTaxon=Plantae&targetTaxon=Gammarus . Note that the attached interaction was tagged by the Encyclopedia of Life's refutation dataset as suspicious.
|
|
Fixed on our end |
|
In update 2021-05-03, I noticed that: $ curl --silent "https://mangal.io/api/v2/interaction?network_id=1113" | jq '.[] | select(.id== 67171)'
{
"id": 67171,
"node_from": 14654,
"node_to": 14643,
"date": "2006-05-19T00:00:00.000Z",
"direction": "directed",
"type": "herbivory",
"method": "observation",
"attr_id": 1,
"value": 1,
"geom": null,
"public": true,
"network_id": 1113,
"created_at": "2019-02-27T00:25:21.659Z",
"updated_at": "2021-04-29T04:45:44.504Z",
"attribute": {
"id": 1,
"name": "presence/absence",
"description": "Presence or absence of a recorded interaction",
"unit": null,
"created_at": "2019-02-21T21:17:01.496Z",
"updated_at": "2019-02-21T21:17:01.496Z"
}
}On #1 (comment) , the following was observed: {
"id": 67171,
"node_from": 14643,
"node_to": 14654,
"date": "2006-05-19T00:00:00.000Z",
"direction": "directed",
"type": "herbivory",
"method": "observation",
"attr_id": 1,
"value": 1,
"geom": null,
"public": true,
"network_id": 1113,
"created_at": "2019-02-27T00:25:21.659Z",
"updated_at": "2019-02-27T00:25:21.659Z"
}So, it appears that interaction with id 67171 was reversed from:
to:
on or before 2021-04-29T04:45:44.504Z ( 29 April 2021 ) |
|
@tpoisot can you please confirm that: The previous example (i.e., 67171) documents a corrected an interaction claim:
to:
However, it appears that some suspicious interactions remain, e.g., Zostera (eelgrass) preys on Gammarus mucronatus remain.
|
|
Looks like network 1114 has some suspicious predation interactions https://mangal.io/api/v2/interaction?network_id=1114&type=predation - https://mangal.io/api/v2/interaction?id=67195 (crustacean parts preying on Elasmopus) and https://mangal.io/api/v2/interaction?network_id=1112&type=predation https://mangal.io/api/v2/interaction?id=67138 - Zostera (eelgrass) preying on Gammarus mucronatus - [
{
"id": 67134,
"node_from": 14620,
"node_to": 14632,
"date": "2005-04-21T00:00:00.000Z",
"direction": "directed",
"type": "predation",
"method": "observation",
"attr_id": 1,
"value": 1,
"geom": null,
"public": true,
"network_id": 1112,
"created_at": "2019-02-27T00:25:11.934Z",
"updated_at": "2019-02-27T00:25:11.934Z",
"attribute": {
"id": 1,
"name": "presence/absence",
"description": "Presence or absence of a recorded interaction",
"unit": null,
"created_at": "2019-02-21T21:17:01.496Z",
"updated_at": "2019-02-21T21:17:01.496Z"
}
}
] |
|
@tpoisot For your convenience, I've included a list of 48 refuted interaction records as provided by Encyclopedia of Life refutation dataset . fyi @KatjaSchulz |
|
I have sincerely no idea how to read this file |
|
All I can say is:
|
|
@tpoisot thanks for your prompt reply.
if you'd like to learn more about the format of the refutation dataset that EOL uses, you might find some inspiration at https://github.com/globalbioticinteractions/refuted-biotic-interactions-by-eol and issues like globalbioticinteractions/globalbioticinteractions#376 .
Ok, I'll try and add duplicate issues as suggested to reference issues posted elsewhere (e.g., #1 and mangal-interactions/mangal-api#55 ). Question - I noticed that https://github.com/mangal-wg/mangal-datasets has not been updated since Oct 2020, but changes were made to records exposed through the mangal api. What is your procedure to keep https://github.com/mangal-wg/mangal-datasets in sync with the live mangal database / api? |
|
The repo is just for initial upload - we then use the API for corrections when required. FYI our effort is currently on migrating to a new cloud provider and updating the stack, so we won't be very reactive with data modifications. As for the EOL format, I'm sure it's interesting, but we'd probably gain more from a plain English issue. Anyways, the issues are confined to a particular dataset, so we'll manually check it at some point in the future. |
|
Hi everyone! @jhpoelen, interaction directions for network 1112, 1113 and 1114 are now resolved. I might have the time to check them. And @tpoisot, you can read the .tsv.txt file in Julia with: |
|
I figured how to read the file 🙂 it's more the semantics that confused me |
|
I figured I wasn't going to teach you anything in Julia, it was just in case 😉. |
|
@BenMerSci @tpoisot thanks again for your prompt replies. For what it is worth, the header of the refutation data: See also https://github.com/globalbioticinteractions/refuted-biotic-interactions-by-eol/blob/main/globi.json for schema mappings. |
|
ps some network id to each interaction record can be added. Happy to add it if someone asks for it. |
|
For each row in the file, there is only 1 species for the refuted interaction. Do you know what is happening with that? With the network_id do you know by any chance if you would be able to get the interaction id? It would really make the investigation easier 😅. |
|
For your convenience, I've added the mangal related refuted dataset below and attached. refuted_mangal_by_eol_with_header.tsv.txt I see a lot of names associated with the source (left side), and target (right side) of the interaction records. I noticed that there might be some false positive refutation. It looks like the name "Fungal" maps to GBIF's https://www.gbif.org/species/4904189 which was somehow classified as a virus. This classification triggered the refutation rule. Aside from that there's a bunch more. Please hollder if you have more questions.
|
|
So the issue was that the "Fungal" group was classified as a virus - which is maybe a decision from the GBIF backbone. We need to dig into the dataset to see if changing the mapping of this node to - "Fungi" is correct. This doesn't seem too bad, and is not a case of a flipped interaction. As I asked before, please open issues in the Mangal repo. Or as a suggestion, as GLOBI is consuming data, maybe you want to point users to the place where they should be opening issues, since the GLOBI team won't be fixing them, and since we won't be monitoring this repo for issues. |
|
@tpoisot much like your idea to make the endpoint for issues configurable. I agree that having folks contact curators of (aggregate) indexed dataset directly would likely be beneficial. See globalbioticinteractions/globalbioticinteractions#661 . re: the suspicious "Fungal" mapping - yes, this likely a name mapping issue. I've opened a separate issue for that at globalbioticinteractions/globi-taxon-names#8 . Please do note that other refutation records were noted, so there's refuted interaction records beyond those that involve the "Fungal" name. |
|
Yes, im looking into these |
|
As for "Fungal", it's probably "Fungal virus" with the " virus" part having been dropped at some point. It happens on some backone (we had to clean a lot of names because of that recently). |
jhpoelen
pushed a commit
to globalbioticinteractions/globi-taxon-names
that referenced
this issue
May 12, 2021
jhpoelen
pushed a commit
to globalbioticinteractions/globalbioticinteractions
that referenced
this issue
May 12, 2021
|
Thanks @tpoisot for your note on the suspicious Fungal mappings. I've prepared a new release of the GloBI taxon graph 0.3.32 at https://zenodo.org/record/4753955 to remove this undesired mapping of names including "[Ff]ungal [Hh]yphae" and "Fungal spores". Changes should propagate over the next couple of days. |
|
It appears that the name mapping updates related to suspicious "Fungal Hyphae" and "Fungal spores" have propagated and no suspicious viral interactions appear in the indexed mangal interaction networks. Also, the original suspicious interactions e.g., Plants eating Gammarus (an amphipod) https://www.globalbioticinteractions.org/?interactionType=interactsWith&sourceTaxon=Plantae&targetTaxon=Gammarus are no longer being indexed by GloBI in the mangal interaction networks. While suspicious interactions remain (e.g., Green Algae (Chlorophyta) eating Gymnoplea (a copepod), and Chlorophyta eating Crustacea (see attached screenshot)), I am sure that @tpoisot et al. will curate their interaction networks as they see fit. That said, I'd be happy to help facilitate reviewing interaction data when desired. @ddkapan @tpoisot thanks for your efforts in carefully reviewing and curating interaction records.
|
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Hi!
Thanks for helping to make existing biotic interaction data easier to find and access!
I was just looking at your GloBI indexed record at https://www.globalbioticinteractions.org/?interactionType=interactsWith&sourceTaxon=Plantae&targetTaxon=Gammarus and I was wondering about the possibility of a REVERSED interaction where a PLANT ate an ANIMAL. For now I'll assume it is reversed, for future downloads of GLOBI data for a project I am working on it would be great to have this correct at the source.
Thanks,
Durrell Kapan
California Academy of Sciences.
The text was updated successfully, but these errors were encountered: